Uncertainty Analysis and Model Reduction Based Global Optimisation of Distributed Large-scale Systems

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

70 Downloads (Pure)

Abstract

Uncertainty arises in many large-scale distributed industrial systems, needing efficient computational tools. Uncertainty propagation techniques have been developed and applied including power series expansions (PSE) and polynomial chaos expansions (PCE). However, such fast low-order approximate models generate errors and, in general, require prior knowledge about uncertainty distribution. In this work, the recursive projection method (RPM) was adopted to accelerate the computation of steady state solutions of complex large-scale dynamic systems. These accelerated models including uncertainty were subsequently utilised in an efficient Bayesian global optimisation framework. The performance of the proposed robust optimisation framework was demonstrated through an illustrative example: a tubular reactor where an exothermic reaction takes place.
Original languageEnglish
Title of host publication30th European Symposium on Computer Aided Chemical Engineering
PublisherElsevier BV
Volume47
Edition1
ISBN (Print)9780128233771
Publication statusPublished - 1 Sept 2020
Event30th European Symposium on Computer Aided Process Engineering -
Duration: 31 Aug 20202 Sept 2020

Conference

Conference30th European Symposium on Computer Aided Process Engineering
Abbreviated titleESCAPE 30
Period31/08/202/09/20

Fingerprint

Dive into the research topics of 'Uncertainty Analysis and Model Reduction Based Global Optimisation of Distributed Large-scale Systems'. Together they form a unique fingerprint.

Cite this